Issue with SI architecture clone _Galaxy Wars

andykmv

Active member
Joined
Jun 27, 2009
Messages
1,535
Reaction score
23
Location
Kilmore, Vic, Australia
i have been helping out baz with his galaxy wars PCB (space invaders architecture, not the same layout though as midway or taito) and i have been looking for pics of symptoms as shown on the SI monitor of typical or unusual faults (such as has been done for pacman repair logs/symptoms!)

or related posts see:
- SI typical issues/pics - http://forums.arcade-museum.com/showpost.php?p=1616786&postcount=1
- 8080 CPU signals etc - http://forums.arcade-museum.com/showpost.php?p=1617698&postcount=1

there are no specific schematics or pcb layouts or pinouts for this Artic pcb so i have used the Midway SI L shaped schematic as a starting point and have managed to identify key components in the architecture, their placement on the board so i am annotating the midway schematic with GW layout coords as i go, and annotating the schematic with any obvious diffs as i go.

i thought i had seen a thread with some pics of typical faults but cant for the life of me find the darned thing.

i am trying to compile a list of visible symptoms that i can use as an aid to debuggin baz's board...

the real questions is, can anyone point me to a bunch of SI video screens images showing typical fault symptoms ? or can anyone send me copies of their pics? particularly interested at this stage in pics of issues from Off to the SI bootup before the game attract screen loads

i have an SI multigame board and a good cpu, just need to work my way thru progressively! i think that once we can get it to boot to the attract mode the going will get easier.



repair progress:

firstly we started off with horizontal stripes on screen, with a broken up lines pattern, then the board died leaving only a white screen.

i have since
- desoldered the cpu and installed a socket. cpu is dud. good one installed.
- removed the rams, installed sockets and replaced rams with rams off another clone board.

the pattern i now see on screen is as follows:
Video%20after%20Ram%20Sockets%20installed%20a.JPG


from some further probing it looks like ram address A07 is stuck so i need to pull out the DSO and check in more detail and poss replace one of the address counters.
 
Last edited:
will have a look and see if i can find the screen fault pics.

in meantime im making my own to go into my space invaders web site covering all the boards plus bootlegs and clones.
 
Could be bad ROM sockets but most likely bad RAM.....are they all soldered in place like on the SI boards? The SI has a power on reset on one of the pins....without it, it will be stuck with a screen like yours is showing. Might be worth checking.
 
i have an SI multigame board and a good cpu, just need to work my way thru progressively! i think that once we can get it to boot to the attract mode the going will get easier.

Have you tried it with the multigame board? It bypasses the roms, usually if you
have clock and power, the multigame board will run and you can see it running
the ram tests. For ram failures the SI kit uses the missile sound to beep out
the code, not sure what what would be wired to on your board. IO PORT 3.
 
thanks for the tips there guys; something i can do then whilst waiting for parts

- pwronreset - still unclear how this is wired up on this pcb so more investigation needed. there is a 555 timer on main brd and 556 on snd brd so may still need snd brd wired up for it to wk properly (if pwronreset line goes to snd brd and then back to cpu). initial aatempts to trace reset line backwards failed. have to try again now that i know more about the chip functions when compared to midway si schematic
Note: the cpu reset is LOW and remains LOW during and after power on, unlike my taito 3 layer si brd.

- we replaced the cpu (original faulty, tested in a midway si)

- i have just finished installing ram sockets and reinstalled ram. now have the screen u see compared to white screen

- i have tried it with the wkng cpu and the si multiboard with wrking cpu, same screen.

- i have clocks but they may be at diff freq's (DMM hz test shows diff results but it could be crap at digital hz measurement). need to make some space to setup the DSO and measure the clocks properly.
i still need to do some detailed investigation into the circuitry around the 7442 and attached ics early in the clock/address chain, to see that they are working correctly

- sound brd not yet connected - awaiting wiring diag from baz, after which i can hook it up and listen for beep codes. on the good side, the 8 audio frequency lines shown in midway diag are generating the correct freqs (well, near enough accord to dmm hz setting)

- awaiting an order for a set of ttl chips for the brd so i can replace poss bad address counter ic

- i still need to investigate the ram output buffers and data bus buffers

at least some progress is made so far from dead to some lines on screen!
 
omegaman, your post also reminded me that the CPU data lines looked to be in HiZ state...so maybe cpu is halted/frozen (its a good cpu)
 
omegaman, your post also reminded me that the CPU data lines looked to be in HiZ state...so maybe cpu is halted/frozen (its a good cpu)

Seems like back when I was messing with SI stuff that the 8080 is much more
sensitive or needy in terms of proper reset. If its not going through the proper
reset sequence possibly by other hardware holding it in reset correctly and
releasing it, that maybe it wont run. I just remember the 8080 being more
critical about this than say z80 or 6502.
 
Hi Andy-

My Taito PCB had a symptom of not booting, but going to the 'test' mode regardless if the ROMs were installed or not.

The line highlighted in green was shimmery and the others were "broken" as seen here. The lines on the other "page" of the test were ok.

It turned out to be the bottom RAM on the left shown here (desoldered).

I'll confirm the RAM position.

SI_RAM1.jpg


PCB1.jpg
 
now wondering if andy has attempted to plug in the braze multi chip into the cpu socket to see if she does anything ?

yep did that, same pattern as shown up top, so i have been looking at the data bus, ready/halt signals etc to see why cpu may be in "halt" type mode where the databus in in HiZ mode.

gotta make some room on the bench to setup your good taito boardset to analyse the signals on that for comparison, and i think my logic probe has crapped out. :(
 
Last edited:
the pattern i see on screen is as follows:
Video%20after%20Ram%20Sockets%20installed%20a.JPG


i am using a normal tv with an rf modulator to take the B/W composite video from the board into the tv

i will shrink the horizontal width as there should prob be more double lines than are seen here

also, still waiting on a slew of parts from futurlec. hasnt shipped yet grr but shouldnt be long.
 
well now we know 100% with accuracy this board is identical to universals cosmic monsters space invaders clone.

anyone know where i can get cosmic monsters schematics ?
 
Back
Top Bottom